Application of machine learning and complex network measures to an EEG dataset from ayahuasca experiments
PLoS ONE December 16, 2022 Caroline L. Alves, Rubens Gisbert Cury, Kirstin Roster et al. 13 citations
Ayahuasca, an Amazonian plant blend used in traditional medicine for centuries, is a promising therapy for neurological and mental diseases. Using an EEG dataset, machine learning and complex network analysis automatically detected changes in brain activity at three data abstraction levels. Connectivity changes between brain regions (correlation of EEG time series) yielded the highest accuracy (92%), followed by raw EEG (88%) and complex network measures (83%). The frontal and temporal lobes were most activated, consistent with prior work. A novel finding identified F3 and PO4 as the most important brain connections, possibly linked to face-recognition-like cognitive processes during visual hallucinations.